../lessons/. Source: semantic-capabilities-reference.md.Semantic Capabilities Cheat Sheet
14 capabilities in 4 domains, plus a step-by-step classification procedure and the tie-breaker rules.
Classification procedure
A raw observation is often a compound statement. Before classifying anything, split it so each piece has exactly one thing it's fundamentally about. Don't force multiple facts under one label just because they came from the same sentence.
Run it through this checklist in order and stop at the first match — the order exists specifically to resolve the tie-breakers below, so don't skip ahead.
Work down this table from top to bottom and stop at the first row whose question matches — don't skip ahead, and don't keep looking once you have a match. The order isn't arbitrary: it's sequenced specifically so the framework's documented tie-breakers resolve themselves without backtracking, for example:
- Row 4 (Provenance) checks for a contrasted information source before row 5 (Observation Point) defaults to Observation Point.
- Row 9 (Infrastructure Output) checks whether the statement survives a service failure before row 10 (Service Outcome) defaults to Service Outcome.
The Companion column flags a capability worth attaching alongside the Primary at that step — see the companion-pairings table below for the fuller picture.
| Capability | Question | Companion |
|---|---|---|
| Asset Management Context | Pure administration — ownership, contract, warranty, lifecycle, not a measurement at all? | — |
| Network Operability | About the current communication/connectivity state itself — reachable right now, latency, signal quality? | — |
| Fallback Behaviour | About what the asset does autonomously because supervisory control or comms was lost? | Network Operability, if the cause is stated |
| Provenance | Specifically contrasts who/what supplied the information — device-reported vs. manually entered vs. externally supplied? | Observation Point, if a location is also stated |
| Observation Point | About where the referent physically or logically sits — a position, not an actor? | Provenance, if the source is worth noting |
| Measurement Quality | About how reliable, accurate, or complete the value is? | — |
| Operational Context | About a dynamic, situational condition that changes moment to moment? | — |
| Physical Context | About a static, structural, fixed feature of the place? | — |
| Infrastructure Output | Would the statement still be true even if the intended service completely failed? | — |
| Service Outcome | Otherwise: is this the real-world effect itself, the "so what"? | — |
| Resource Consumption | About cost or resources spent — energy, water, fuel, battery? | — |
| Observation Scope | About what portion or coverage of the system the value represents, including one number combining multiple assets or locations? | Observation Method, if the aggregation technique matters |
| Temporal Semantics | About what time span the value covers, including one number combining multiple readings over time (e.g. a rolling average)? | — |
| Observation Method | Otherwise: about the technique that produced the value (measured / estimated / predicted / simulated / proxy)? | — |
Common companion pairings
Primary is always exactly one capability. These pairings come up often enough to check for by habit:
| Primary | Common companion | Why |
|---|---|---|
| Fallback Behaviour | Network Operability | The connectivity loss is usually why the fallback triggered. |
| Provenance | Observation Point | A device name can answer both "who" and "where" at once. |
| Observation Point | Provenance | Same pairing, reversed — when position is the main point but the source is still worth noting. |
| Observation Scope | Observation Method | Recording how a multi-asset aggregate was computed, not just that it covers many assets. |
| Provenance | Observation Method | A model-derived or predicted value usually needs both — who/what produced it, and how. |
| Service Outcome / Resource Consumption | Temporal Semantics | A cost or outcome figure often carries an implicit window ("today," "this cycle"). |
Domain Semantics — what the municipality ultimately cares about
| Capability | Question | Examples |
|---|---|---|
| Service Outcome | Did the real-world effect happen? | Street illuminance at road level; soil moisture at root level; delivered water pressure |
| Infrastructure Output | What did the asset itself produce? | Lumens emitted; pump flow rate; valve opening %; motor speed |
| Resource Consumption | What did it cost to produce that? | Energy; water; fuel; battery; active/reactive power; voltage; frequency |
Observation Semantics — how it was observed
| Capability | Question | Examples |
|---|---|---|
| Observation Point | Where does the referent physically/logically sit? | Luminaire; cabinet; pump; valve; pipeline; root zone; weather station |
| Observation Scope | What portion of the system does this cover? | Single asset; group; street; irrigation zone; district; municipality |
| Observation Method | What technique produced this value? | Measured; estimated; inferred; predicted; aggregated; simulated; model-derived; proxy |
| Temporal Semantics | What time span does this represent? | Instantaneous; interval; sampling period; aggregation window; historical; forecast; horizon |
Interpretation Semantics — how to understand and trust it
| Capability | Question | Examples |
|---|---|---|
| Provenance | Which system/actor supplied this? | Device sensor; control cabinet; weather service; GIS; human operator; Digital Twin; predictive model |
| Measurement Quality | How reliable/accurate is it? | Accuracy; precision; confidence; completeness; consistency; availability; resolution; uncertainty |
| Operational Context | What dynamic condition modulates it now? | Weather; rainfall; humidity; temperature; wind; traffic; occupancy; season |
| Physical Context | What static feature of the place affects it? | Buildings; trees; vegetation; terrain; slopes; orientation; shadows; surface materials |
Operational Semantics — how it's managed and kept running
| Capability | Question | Examples |
|---|---|---|
| Asset Management Context | Who owns/maintains it, what lifecycle stage? | Ownership; org; contracts; warranty; lifecycle stage; install date; remaining useful life |
| Network Operability | Is it reachable right now, how well? | Availability; connectivity status; latency; packet loss; signal quality; comms technology |
| Fallback Behaviour | What does it do autonomously if control is lost? | Local schedules; autonomous control; safe/emergency mode; manual override |
Tie-breakers — the calls that feel like they could go two ways
Would the sub-observation still be true even if the service completely failed?
- Yes → Infrastructure Output
- No, it's the effect itself → Service Outcome
What got combined?
- Multiple places → Observation Scope (Observation Method as companion)
- Multiple readings over time → Temporal Semantics
- Default → Observation Point (a position), when a device name could answer either
- Flip to Provenance only when the sub-observation specifically contrasts sources — device-reported vs. manually entered vs. externally supplied
Would this still be true a year from now regardless of today's conditions?
- Yes, fixed → Physical Context
- Depends on today → Operational Context
- Network Operability is the connection state itself
- Fallback Behaviour is what the asset autonomously does because of that state (usually carries Network Operability as a companion explaining why)
Worked example
| Sub-observation | Step matched | Primary | Companion |
|---|---|---|---|
| Owned by Parks Dept, 5-year contract | 1 | Asset Management Context | — |
| Signal quality degraded, connection lost | 2 | Network Operability | — |
| Now running local emergency schedule | 3 | Fallback Behaviour | Network Operability |